Vellum - HumanEval: leaderboard
Vellum's independently-run HumanEval coding evaluation. Function completion tasks measuring code generation ability.
Metric: Pass@1 (%). Source: www.vellum.ai. Saturation forecast: Around January 2027. 47 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 96.2 | #19 |
| 2 | Claude Mythos 5 | 95.5 | #2 |
| 3 | Claude Fable 5 | 95 | #16 |
| 4 | GPT-5.6 Luna | 93 | #68 |
| 5 | Claude Opus 4.8 | 88.6 | #37 |
| 6 | Claude Opus 4.7 | 87.6 | #49 |
| 7 | Claude Sonnet 5 | 85.2 | #57 |
| 8 | Claude Sonnet 4.5 | 82 | #142 |
| 9 | Claude Opus 4.5 | 80.9 | #83 |
| 10 | Claude Opus 4.6 | 80.8 | #63 |
| 11 | Gemini 3.1 Pro (Preview) | 80.6 | #58 |
| 12 | DeepSeek V4 Pro | 80.6 | #102 |
| 13 | MiniMax-M3 | 80.5 | #136 |
| 14 | Kimi K2.6 | 80.2 | #104 |
| 15 | GPT-5.2 | 80 | #110 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vellum-humaneval · How It Works · Data refreshed daily, snapshot 2026-10-20.